Distributed static compensator (DSTATCOM) for voltage support in single wire earth return (SWER) networks

Distributed static compensator (DSTATCOM) for voltage support in single wire earth return (SWER) networks

This investigation is concerned with the effectiveness of Distributed STATic COMpensators (DSTATCOMs) at providing voltage support in Single Wire Earth Return (SWER) networks. The reason for the focus on SWER lines is the high cost of upgrading them in the traditional way to solve voltage regulation problems that result from load growth in some of the feeders. A number of aspects of DSTATCOM installation and operation have been explored. These include their location, reactive power circulation, reactive power prioritising, four quadrant operation and the timing of installation and operation.

It has been possible to derive analytical expressions only for the case of a single Thevenin source equivalent and a single load in parallel with a DSTATCOM. From one of those expressions it was deduced that, on a voltage increment per kVAr basis, DSTATCOMs are most effective as voltage regulators when they are installed at the customer terminals rather than further upstream into the network. This result has been found to apply generally to all practical SWER lines. Another derived expression predicted a peak value of customer terminal voltage when active power (P) and reactive power (Q) are injected by the DSTATCOM at constant kVA (S). This maximum voltage represents a stability limit for the case where DSTATCOMs are controlled to operate at constant kVA. Load flow studies revealed that in general this stability limit exists for all practical SWER lines.

To avoid VAr circulation it is proposed that droop control with hysteretic band is used for DSTATCOM operation. The standard Newton-Raphson load flow formulation has been extended to accommodate DSTATCOMs operating under droop control and with operating point on a defined trajectory on the P-Q plane. Four defined trajectories have been investigated under given hourly load demand profile. These are the Q-only scheme, the constant kVA scheme with Q-priority, the load power factor follow scheme and the power factor correction scheme. For each one of those defined trajectories a modified Jacobian had to be derived. Load flow studies were based on each of the modified formulations. The load flow programs were designed to automatically provide a solution for each hour of a 24-hour demand profile representing the worst case peak demand for a particular year in the life of any practical SWER line. The customer DSTATCOM is either left on line, brought on line, left off line or taken off line, depending on the calculated customer voltage. Those special features of the load flow programs allowed them to be used to determine when and at what customer location DSTATCOMs should be installed and what their ratings should be. While the focus of the thesis has been on undervoltage problems; the proposed solutions and algorithms are applicable to overvoltage problems caused by the Ferranti Effect.